Method of applying a catalyst to a support

This experiment shows that platinum can be made to reduce on the surface of alumina. Five grams of gamma alumina powder was slurried in about 25 grams of water. A solution was prepared that contained 0.075 grams of platinum as H 2 PtCl 6 and 0.0404 grams of hydrazine dihydrochloride, NH 2 NH 2 2HCl, which is just enough to reduce all of the platinum to the metal. The weight of this solution was about 10 grams. The slurry of alumina was stirred while the solution was added a drop at a time. Reduction occurred immediately as the drops fell upon the alumina, as shown by the alumina changing from white to black. The solution that contained the H 2 PtCl 6 and the NH 2 NH 2 .2HCl was transparent pale yellow. This solution remained clear and transparent until it fell upon the alumina. That is, there was no reduction until the solution met the alumina.

The experiment of Example 1 was repeated with other surfaces substituted for the gamma alumina:

______________________________________

Substrate Surface Area (m.sup.2 /g)

Result

______________________________________

Silica 200 No reduction

Titanium oxide

unknown No reduction

Cerium oxide

128 No reduction

Carbon 1020 Complete reduction

______________________________________

The carbon, of course, was black at the beginning. But the color of the solution turned from yellow to colorless, indicating complete reduction.

The experiment of Example 1 was repeated with the difference that the solution of H 2 PtCl 6 and NH 2 NH 2 .2HCl contained only half of the original amount of NH 2 NH 2 .2HCl. That is, it contained only enough NH 2 NH 2 .2HCl to reduce half of the platinum. The alumina darkened slowly and the final color was gray, not black. When the slurry was allowed to settle, the supernatant solution was pale yellow, indicating the presence of unreduced platinum. When more NH 2 NH 2 .2HCl was added, the alumina darkened and the yellow color disappeared. This experiment confirms, again, that the reduction occurs on the surface of the alumina.

The experiment in Example 1 was repeated with the difference that the solution of H 2 PtCl 6 and of NH 2 .2HCl were not mixed together, but instead they were added separately to the slurry of alumina. The result was the same whether the H 2 PtCl 6 or the NH 2 NH 2 .2HCl was added first. The alumina did not begin to darken until about five minutes after the second solution had been added. Then the alumina darkened gradually over a period of about two minutes. The final color of the alumina was light gray instead of black. This suggests that the platinum had time to distribute itself over the area within the particles of alumina powder before the reduction started.

The purpose of this experiment was to confirm the observation of Example 4, that when the solutions of H 2 PtCl 6 and NH 2 NH 2 .2HCl are added to the alumina separately, the platinum has time to distribute itself inside of the alumina particles. The alumina was in the form of beads having a diameter of about 3 mm, instead of a powder. The beads could not form a slurry, but instead they were covered with water. When the solution of H 2 PtCl 6 was added first, and later the NH 2 NH 2 .2HCl was added separately, the beads did not start to darken for about 15 minutes. When the darkening was complete, the beads were split. There was a darkened surface shell having a thickness of about one tenth of the diameter of a bead.

When the solutions were mixed and then added to the beads, the beads started to darken within 12 seconds. Splitting the beads revealed a surface shell that had no perceptible thickness at all. The platinum was confined to the outermost surface of the bead.

The purpose of this experiment was to demonstrate the reduction of palladium on the surface of alumina. Three grams of powdered gamma alumina was slurried in about 20 grams of water. A solution was prepared that contained 0.045 g of palladium as H 2 PdCl 4 and 0.0221 g NH 2 NH 2 .2HCl, which is just enough to reduce all of the palladium to the metal. The weight of this solution was about 10 grams. The slurry of alumina was stirred and the solution was added, a drop at a time. Reduction started when the first drops fell upon the alumina.

The experiment of Example 6 was repeated with the difference that the solution of H 2 PdCl 4 was added first to the slurry of alumina, followed by the solution of NH 2 NH 2 .2HCl. The alumina began to darken within about one second after the addition of NH 2 NH 2 .2HCl started.

Platinum of valence 4 was reduced to platinum of valence 2 by the reaction

H.sub.2 PtCl.sub.6 +(1/2)NH.sub.2 NH.sub.2. 2HCl→H.sub.2 PtCl.sub.4 +(1/2)N.sub.2 +3HCl

The H 2 PtCl 6 was in a concentrated solution that contained about 5%, by weight, of platinum. The NH 2 NH 2 .2HCl was added as the solid salt, a little at a time. Reduction is complete when the evolution of nitrogen gas stops.

The previous examples cited solutions that contained both H 2 PtCl 6 and NH 2 NH 2 .2HCl, and there was no reduction of the platinum in those solutions. The reason is that reduction occurs only at the much higher concentration in this Example 8.

The purpose of this experiment was to demonstrate the reduction of platinum from H 2 PtCl 4 . Ten grams of gamma alumina powder was slurried in about 40 grams of water. A solution was prepared that contained 0.105 grams of platinum, as H 2 PtCl 4 . Another solution contained 0.030 grams of NH 2 NH 2 .2HCl. The slurry of alumina was stirred. Each solution was added a drop at a time. The H 2 PtCl 4 was added first. Reduction started when the first drops of NH 2 NH 2 .2HCl solution fell upon the alumina.

The experiment was repeated and the solutions of H 2 PtCl 4 and NH 2 NH 2 .2HCl were mixed together. The mixed solution weighed about 10 grams. When this solution was added to the slurry of alumina, drop by drop, reduction started when the first drops fell upon the alumina.

The purpose of this experiment was to scale up the reduction demonstrated in Example 9, and to produce a catalyst coating whose activity could be tested in a later example. The alumina powder used here was made by calcining Disperal, which is an alpha monohydrate, Al 2 O 3 .H 2 O, that contains additional water of hydration. It is supplied by Condea Chemie GmbH, of West Germany. The gamma alumina that was impregnated with platinum was made by calcining Disperal at 780° C. for two hours. The weight loss upon calcining was 22.8%, which indicates that all of the water was evolved. The product is gamma alumina having a surface area of about 125 m 2 /g. Fifty grams of this gamma alumina was slurried in about 200 g of water. The slurry was stirred while a solution of H 2 PtCl 4 that contained 0.52 g of platinum was added a drop at a time. Then a solution that contained 0.15 g of NH 2 NH 2 .2HCl was added a drop at a time. This is about 110% of the weight of NH 2 NH 2 .2HCl needed to reduce all of the platinum. The alumina started to turn black within seconds after starting the addition of the NH 2 NH 2 .2HCl. After all of the NH 2 NH 2 .2HCl had been added, and the reduction was complete, ammonium hydroxide was added to raise the pH of the slurry to about 8. Raising the pH causes the alumina to settle rapidly and leave a supernatant solution that is clear and not fogged by peptized alumina. This also speeds the filtration in the next step.

The alumina, now impregnated with 1.04% of platinum, was washed on a filter until the filtrate was free of chloride ion, as shown by the absence of a precipitate when silver nitrate was added to the filtrate. The wet filter cake weighed 102 g. It was charged directly to a ball mill without drying. The filter cake that was charged to the ball mill was assumed to contain 48 g of alumina, to allow for a 2 g loss during filtering and recovering the cake from the filter.

The charge to the ball mill had the following compositon:

______________________________________

Weight, grams Component

______________________________________

48.0 Assumed recovery of the

starting alumina

17.3 Alumina made by calcining

Disperal at 400° C.

20.9 Cerium oxide

5.7 Concentrated (70%) nitric

acid

53 Water in addition to the

water in the filter cake

______________________________________

The alumina made by calcining Disperal at 400° C. is not completely dehydrated. The weight loss upon calcining was 9.5%, as compared with a 22.8% loss upon complete dehydration at 780° C. The incompletely dehydrated alumina is partly peptized by the nitric acid, and thereby forms the "glue" that makes this catalyst coating adhere to metal.

The cerium oxide is a component of a three-way automobile catalyst. The composition of the exhaust gas alternates between an excess of oxidants, O 2 and NO x , and an excess of reductants, hydrocarbons, and CO. The cerium alternates between Ce 2 O 3 and CeO 2 , whereby it stores up oxygen during the oxidizing half cycle and releases this oxygen during the reducing half cycle. The cerium oxide used here has an area of about 125 m 2 /g.

This charge was ball milled for 4.7 hours. The product is a washcoat that forms the catalyst coating of this invention. It is a stable suspension; some solids will settle out over a period of weeks, but no faster than that. The specific gravity is about 1.6. The percentage, by weight, of solids that remain after drying and calcining is about 45%. The adherence of the catalyst coating was tested in the following way. A strip of metal foil having a thickness of 0.0025 inches and having the composition of U.S. Pat. No. 4,601,999 was coated with washcoat. Four coatings were applied with a brush. Each coating was dried and calcined before the next coating was applied. After the fourth coating, the loading of calcined catalyst coating on the strip was 49 milligrams per square inch. The strip was penetrated with fine shot from a 0.22 caliber shot cartridge. The coating adhered to the strip right up to the periphery of the shot holes, which is an indication of good adherence.

The purpose of this example is to demonstrate the reduction of rhodium on the surface of alumina. This example includes three experiments. In each experiment, 3.0 g of gamma alumina (Disperal calcined at 780° C.) was slurried in about 20 g of water. A solution of RhCl 3 that contained 0.0312 g of rhodium was added with mixing. Then a solution containing 0.0239 g NH 2 NH 2 .2HCl was added. This is just enough NH 2 NH 2 .2HCl to reduce all of the metal. Unlike the case of platinum or palladium, the reduction of rhodium does not start immediately, but instead it needs some urging. The slurry was warmed slowly while mixing. Then NH 4 OH was added to raise the pH. The only difference among the three experiments was the temperature at which NH 4 OH was added to raise the pH.

Here are the results:

______________________________________

Temperature at which

NH.sub.4 OH was added (°C.)

Result

______________________________________

40 Started to darken at 47° C.,

darkening complete at 54° C.

50 Darkened immediately

62 Darkened immediately

______________________________________

The purpose of this experiment was to scale up the reduction of rhodium demonstrated in Example 11, and to produce a catalyst coating whose activity could be tested in a later example. The alumina powder that was impregnated with rhodium was the same as in Example 10, namely Disperal calcined at 780° C. Fifty grams of this alumina was slurried in about 200 grams of water. A solution of RhCl 3 that contained 0.52 g of rhodium was added, followed by a solution that contained 0.40 g of NH 2 NH 2 .2HCl. This is just enough NH 2 NH 2 .2HCl to reduce all of the rhodium. The slurry was heated and stirred. Here are the results:

______________________________________

Minutes of Heating

Temp (°C.)

Observation

______________________________________

0 25

4 35

7 48

9 56 The red color of

RhCl.sub.3 is turning

to yellow

12 61 Darkening slowly

13 64 Almost completely

darkened

14 66 Darkening complete

______________________________________

There was no need to raise the pH. The reduction occured anyway. The alumina, now impregnated with 1.04% of rhodium, was washed free of chloride ion, and charged to a ball as in Example 10. The charge to the ball mill was:

______________________________________

Weight, grams Component

______________________________________

48.0 Assumed recovery of the

starting alumina

17.3 Alumina made by calcining

Disperal at 400° C.

20.9 Cerium oxide

5.7 Concentrated nitric acid

50 Water in addition to the

water in the filter cake

______________________________________

The milling time was 4.1 hours. The adherence of the washcoat was proved by the method of Example 10.

The purpose of this experiment was to scale up the reduction of palladium demonstrated in Example 7, and to produce a catalyst coating whose activity could be tested in a later example. The alumina that was impregnated with palladium was the same as in Examples 10 and 12, Disperal calcined at 780° C. Fifty grams of this alumina was slurried in about 200 grams of water. A solution of H 2 PdCl 4 that contained 0.52 grams of palladium was added, followed by a solution that contained 0.27 grams of NH 2 NH 2 .2HCl. This is 105% of the amount needed to reduce all of the palladium. The NH 2 NH 2 .2HCl solution was added with a dropper. The reduction started as soon as the first drops of NH 2 NH 2 .2HCl fell upon the alumina. The alumina, now impregnated with 1.04% of palladium, was washed and charged to a ball mill as in Example 10 and 12. The charge to the ball mill was the same as before. The adherence of this washcoat to a metal strip was proved by the method of Example 10.

This experiment was almost a duplicate of Example 10. The only difference was that NaHCO 3 was added to the slurry of alumina to react with the HCl that is produced in the reduction step. As before, 50 grams of alumina was slurried in about 200 grams of water. Then 1.88 grams of NaHCO 3 was added to the slurry. This is 105% of the amount needed to neutralize the HCl. Then the solutions of H 2 PtCl 4 and of NH 2 NH 2 .2HCl were added as in Example 10. The reduction started when the first drops of NH 2 NH 2 .2HCl fell upon the alumina. After the reduction, the pH was about 7, as compared with about 3-4 in the absence of NaHCO 3 .

The charge to the ball mill was the same as in Example 10. The milling time was 4.7 hours. The adherence of the washcoat was proved by the method of Example 10.

Prior to the present invention, my preferred method for applying a catalyst coating to a metal support was to coat the support in two steps. First, the support was coated with a washcoat of alumina that contained no catalyst metal. Then, the alumina was impregnated with platinum from a solution of [Pt(NH 3 ) 4 ](OH) 2 , platinum tetrammine dihydroxide. To make this platinum precursor, H 2 PtCl 4 is converted to [Pt(NH 3 ) 4 ]Cl 2 , and then the chloride ions are replaced by hydroxide ions by ion exchange. This platinum precursor is expensive, but it decomposes upon heating without leaving any residue in the alumina coating, so that the catalyst coating has high activity. The purpose of this Example 15 was to impregnate alumina powder with this pretested platinum precursor, and to make a washcoat that would serve as a a standard of comparison for the other washcoats produced by the method of this invention.

The alumina that was impregnated with platinum was the same as in Example 10. Fifty grams of this alumina was wetted with 55 grams of solution that contained 0.52 grams of platinum as [Pt(NH 3 ) 4 ](OH) 2 . This was just enough solution to wet the alumina. The alumina was dried and calcined at 400° C. The platinum precursor decomposed and what remained of the precursor was platinum in the form of the metal or the oxide. The calcined alumina was light gray and weighed 50.5 g. The charge to the ball mill was:

______________________________________

50.5 g impregnated alumina

18.0 alumina made by calcining Disperal at 400° C.

21.8 cerium oxide

6.0 concentrated nitric acid

98 water

______________________________________

The charge was milled for 4.0 hours. The adherence of this washcoat was proved by the method of Example 10.

The purpose of this experiment was to scale up the reduction of platinum demonstrated in Example 4, and to produce a catalyst coating whose activity could be tested in a later example. The alumina powder that was impregnated with platinum was the same as in Example 10. Fifty grams of this alumina was slurried in about 200 grams of water. A solution of H 2 PtCl 6 that contained 0.52 grams of platinum was added with a dropper, followed by a solution that contained 0.30 grams of NH 2 NH 2 .2HCl. This is 107% of the amount needed to reduce all of the platinum. After all additions, the temperature of the slurry was 22° C. The slurry started to darken after 8 minutes, and then darkened rapidly after 10 minutes of stirring. The charge to the ball mill was the same as in Example 10. The milling time was 4.0 hours. The adherence of this washcoat was proved by the method of Example 10.

The catalyst coatings were tested by painting them onto metal foil having the composition disclosed in U.S. Pat. No. 4,601,999. The strips were corrugated so they could be formed into honeycombs. For these tests, the honeycombs had a length of 0.5 inches, and a diameter of 0.75 inches. An important parameter of the test honeycomb is the loading of catalyst metal, in grams per cubic foot. The loading is controlled by the percentage, by weight, of catalyst metal in the coating, and the weight, per unit area, of the catalyst coating on the metal foil. The coatings made in the foregoing examples were calculated to give a metal loading of 20 grams per cubic foot.

The test results are given in Table I. Table I shows the percentages of hydrocarbons, carbon monoxide, and nitric oxide, respectively, that were converted into carbon dioxide or water, using a catalytic converter made from each sample. The two control samples were made by the technique of the prior art, i.e. by first coating the strip with alumina, calcining the strip, and then impregnating the strip with a solution of platinum tetrammine. The samples made according to the invention were prepared by applying the washcoat which already contained the catalyst metal.

Table I shows that:

1. Palladium in Example 13 has the highest activity, higher than Control II, with more than twice the metal loading.

2. The most active platinum catalyst is made from H 2 PtCl 4 , in Examples 10 and 14.

3. Catalysts made with H 2 PtCl 4 are more active than Control I, with the same metal loading.

__________________________________________________________________________

(Example 17)

Testing the Activity of the Catalyst Coatings

Percent Conversion

Coating from

Catalyst

Loading of Catalyst

R = 1.67 R = 0.97 R = 0.67

This experiment first showed that hydrazine, NH 2 NH 2 , reduces platinum just as effectively as the dihydrochloride. Hydrazine is preferred because it introduces less chloride ion that must be washed out. The alumina used here was made by calcining Disperal at 500° C. The weight loss was 11.9%. The area was about 150 m 2 /g. Eighteen grams of this alumina was slurried in about 100 grams of water. A solution containing 0.19 grams of platinum, as H 2 PtCl 6 , was added, followed by a solution that contained 0.033 grams of NH 2 NH 2 . This is 107% of the weight needed to reduce the platinum. The temperature of the slurry was 22° C. The reduction was complete after 2.5 minutes. Separate tests showed that hydrazine is effective for reducing rhodium and palladium.

This is an eight-fold scale up of the production level of Example 10. The weight of alumina powder was 400 grams. It was slurried in about 2000 grams of water. The slurry was stirred rapidly, while a solution that contained 4.16 grams of platinum as H 2 PtCl 4 was added during a period of about 13 minutes. The volume of the solution was about 500 ml. The solution was more dilute than in Example 10, and the time for adding it was longer. The purpose was to avoid even momentary pockets of high concentration of platinum in the solution that contacts the alumina, so that the platinum would be dispersed uniformly over all of the particles of the alumina powder. The platinum was reduced with hydrazine instead of the dihydrochloride. Then the pH was raised to about 9 with NH 4 OH and the slurry was filtered. The alumina was washed on the filter until the filtrate was free of chloride ion. The wet filter cake was charged directly to the ball mill, as before. The charge to the ball mill was:

______________________________________

398 grams Assumed recovery of the starting alumina

143 Alumina made by calcining Disperal at 500° C.

174 Cerium oxide

48 Concentrated (70%) nitric acid

393 Water in addition to the water in the

filter cake

______________________________________

The charge was ball milled for six hours. The adherence of the washcoat was proved by the method of Example 10.

This is a two-fold scale up of the production level of Example 12. The weight of alumina powder was 100 grams. It was slurried in about 400 ml of water. A solution that contained 1.04 grams of rhodium, as RhCl 3 , was added slowly with a dropper while the alumina slurry was stirred rapidly. The solution of RhCl 3 weighed 52 grams. The rhodium was reduced with 0.12 grams of hydrazine, which is just equal to the theoretical amount for complete reduction. The slurry was warmed after the hydrazine was added. Here are the results:

______________________________________

Minutes of Heating

Temp. (°C.)

pH Observation

______________________________________

2 22 6.2

6 31 6.2

8 36 5.7

12 46 5.4

13.5 50 5.3 Added NH.sub.4 OH to

raise the pH

15 54 8.4 Slurry is

darkening

17 56 8.1

19 56 7.8

21 59 8.0 Darkening is

complete

______________________________________

Three 100-gram batches of alumina and one 50-gram batch were impregnated with rhodium in this way. The four batches of impregnated alumina were combined and washed on the filter as before. The charge to the ball mill was:

______________________________________

348 grams Assumed recovery of the starting alumina

125 Alumina made by calcining Disperal at 500° C.

153 Cerium oxide

42 Concentrated (70%) nitric acid

342 Water in addition to the water in the

filter cake

______________________________________

The charge was milled for five hours. The adherence of the washcoat was proved by the method of Example 10.

The washcoats of Examples 19 and 20 were tested by the method of Example 17. The test results are given in Table II. Control III was made in the same way as Controls I and II. The symbols in Table II have the same meaning as in Table I. The "pulsator" mentioned in Table II is an apparatus for aging catalyst in an atmosphere that simulates automobile exhaust. The simulated exhaust contains the expected poisons such as lead and phosphorous.

__________________________________________________________________________

(Example 21)

Testing the Activity of the Catalyst Coatings

Percent Conversion

Coating from

Conditioning of

R = 1.7 R = 0.97 R = 0.70